This is very similar behaviour to that that i have reported this morning - weird things happening when hold it part of the equation.

I have been having the same problems. Looking back it has been going on for over a month. And it cannot be explained by low temperatures. - last night didn’t go below 7°.

See thread ‘Growatt Inverter errors’.

terrible-guidance

    kerregan Looks like we didn't get a successful Growatt API response back when trying to set the instructions.

      A quick update: -
      Last night I set a charge window with hold turned off, and it charged successfully.

      What I'm totally baffled by, is that when I have an afternoon charge window set (with hold turned on) it seems to work fine?!

      Thinking about what useful data I could gather, would setting two overnight charge windows (say 22:00-02:00 and 02:00-06:00) the first with hold turned on, and the second with hold turned off help with troubleshooting?

        merlinsnotin We have fixed a bug where Agile start window with a charge slot start could end up in Hold instruction being sent after charge slot start.

          admin Thanks for the update. I'll turn hold back on for tonight and let you know the outcome tomorrow.

            admin
            Overnight I have been using a ping testing tool to check the connection status of the Growatt dongle. It seems that it goes in to a snooze mode and can take several seconds to wake up after receiving a ping. I have no idea why. Nothing else on my home network seems to have the same issue. It doesn’t seem to happen in the day time.

            I wonder though whether WW could retry, once, after a failure, say after 1 minute. It would be much appreciated by me and you will have seen that I am not the only one who seems to have this problem.

            My log from last night is attached.

            Thanks

            terrible-guidance
            I

              kerregan We already have a retry policy in place. As you can tell from our logs.

              That, combined with the retry delay in 30 mins should cover basic temporary glitches.

              It does sound like yours is a more fundamental comms/networking issue though, given that the 3:00 and 3:30 both failed after 3 retries.

                admin Thanks. The snoozes seem to take several seconds to wake up - one I saw was over 30 seconds so longer than the retry period. Anyway, I’ve set up a stay alive ping to the Growatt dongle to see if that solves the issue. I’ll let you know :)

                  Hold is now a total mess for me.
                  Charge windows with hold on not charging. Also, I had a charge window set with zero slots, which worked as expected. Unfortunately after that hold slot ended, it didn't revert and I continued to run from the grid rather than the battery.

                  Is it just me who is suffering from this?

                    merlinsnotin What does your Giv log tell you? At the end of the hold window? Can you see discharge rate being set back to max?

                    Sorry for posting and vanishing.
                    I looked at the Giv log and it didn't change back. I tried another hold later on and it worked fine, so either there was a random blip, or I messed something up.

                    Since then I've also done some more testing...

                    I set four charge windows with different settings:-
                    a) 01:00-04:00 with 1 slot and hold turned on (03:30-04:00 was the selected charge slot)
                    b) 04:30-06:30 with 4 slots and hold turned off
                    c) 07:00-09:00 with 1 slot and hold turned on (07:00-07:30 was the selected charge slot)
                    d) 09:30-10:30 with 2 slots and hold turned on

                    a) and c) failed to charge, however b) and d) charged successfully.

                    No surprise that b) worked fine, but the fact that d) worked while a) and c) didn't was strange.

                    Looking at the Giv logs, the only difference in the commands received is that d) received the eco set to off and discharge set to 0 before the AC start and end time.

                    Could the order the commands are received in be causing the issue?

                      merlinsnotin Could the order the commands are received in be causing the issue?

                      Shouldn't do, and if it does, how would we know which order it should be in?

                      merlinsnotin Also, I had a charge window set with zero slots, which worked as expected. Unfortunately after that hold slot ended, it didn't revert and I continued to run from the grid rather than the battery.

                      When was this, which window, 2 days ago? I'd like to check out logs at the end of the charge window with hold set and why it was not reverted, just making sure nothing is wrong at our end. Thanks!

                        admin

                        When was this, which window, 2 days ago? I'd like to check out logs at the end of the charge window with hold set and why it was not reverted, just making sure nothing is wrong at our end. Thanks!

                        It would have been Thursday (5th) afternoon expected to end at 4pm. I'm now thinking that I must have set the end time wrong or another PEBCAK type issue.

                        Shouldn't do, and if it does, how would we know which order it should be in?

                        If I'm the only person with a Giv system who's experiencing hold slots not charging, and it's working fine for everyone else, then fair enough. It just seems odd to me that given the same commands were issued in all the test cases, that the only difference between running successfully and failing, is the order that the commands were received in.

                        kerregan

                        admin I spent several days pinging the Growatt dongle, waking in the middle of the night to check the api was alive and generally monitoring. At the end of this, the problem has gone away, even though I have actually changed nothing.

                        For now, I will lay the blame with the Growatt servers.